Smart speakers and similar devices permit users in different locations to speak to each other remotely, using voice only, or video, while also providing a convenient user interface device to an intelligent automated assistant service that can process speech input and carry out simple or sophisticated tasks in response. However, when a user places a call via a smart speaker to a user who is not currently available to take the call, the caller faces the same options as a caller using conventional telephone technology. A caller may either re-attempt the call until the user answers, or else resort to a message sent by a different protocol in order to solicit a call back or to schedule a mutually agreeable time to speak. Either option consumes excess time and network resources.